Archives of Biochemistry and Biophysics 561 (2014) 64–73

Contents lists available at ScienceDirect

Archives of Biochemistry and Biophysics

journal homepage: www.elsevier .com/ locate /yabbi

Review

Bone Morphogenetic Proteins: Structure, biological functionand therapeutic applications

http://dx.doi.org/10.1016/j.abb.2014.07.0110003-9861/� 2014 Published by Elsevier Inc.

⇑ Corresponding author at: Av. Nossa Senhora das Graças, 50. Prédio 6 – Diretoriada DIPRO, 2o Andar. Xerém, Duque de Caxias, 25250-020 RJ, Brazil.

E-mail addresses: ancoc@iq.usp.br (A.C. Carreira), gutemberg.alves@id.uff.br(G.G. Alves), wzambuzzi@ibb.unesp.br (W.F. Zambuzzi), mcsoga@iq.usp.br(M.C. Sogayar), jmgranjeiro@inmetro.gov.br (J.M. Granjeiro).

1 Abbreviations used: BMPs, Bone Morphogenetic Proteins; MSC, mesenchylls; IGFs, Insulin-like Growth Factors; PDGF, platelet-derived growth factbroblast Growth Factors; EGF, Epidermal Growth Factor; TGF-b, Tranrowth Factor Beta; MIS, Mullerian Inhibiting Substance; GDFS, Groifferentiation Factors; Cer1, Cerberus; SOST, Sclerostin; Tsg, Twisted gastMPR-IA, type IA BMP receptor; BMPR-IB, type IB BMP receptor; MAPK,ctivated protein kinase; PK A, Protein Kinase A; SHH, sonic hedgehog; BMceptors; bFGF basic fibroblast growth factor; GM-CSF, granulocyte-malony stimulating factor; BEVS, baculovirus/insect cells system; iPSCs,

luripotent stem cells.

Ana Claudia Carreira a,b, Gutemberg Gomes Alves c, William Fernando Zambuzzi d, Mari Cleide Sogayar a,b,José Mauro Granjeiro e,f,⇑a Chemistry Institute, Biochemistry Department, University of São Paulo, São Paulo 05508-000, Brazilb NUCEL-NETCEM Cell and Molecular Therapy Center, Medical Clinics Department, School of Medicine, University of São Paulo, São Paulo, 05508-000 SP, Brazilc Cell and Molecular Biology Department, Institute of Biology, Fluminense Federal University, Niterói, RJ, Brazild Department of Chemistry and Biochemistry, Biosciences Institute, UNESP: Universidade Estadual Paulista, Botucatu, SP, Brazile Bioengineering Division, National Institute of Metrology, Quality, and Technology, Duque de Caxias, RJ, Brazilf Department of Dental Materials, Dental School, Fluminense Federal University, Niteroi, RJ, Brazil

a r t i c l e i n f o a b s t r a c t

Article history:Received 7 April 2014and in revised form 1 July 2014Available online 17 July 2014

Keywords:Bone Morphogenetic ProteinsOsteogenesis inducing proteinsPeptide growth factorsBone repairExtracellular matrixTissue engineering

Bone Morphogenetic Proteins (BMPs) are multifunctional secreted cytokines, which belong to the TGF-bsuperfamily. These glycoproteins act as a disulfide-linked homo- or heterodimers, being potent regulatorsof bone and cartilage formation and repair, cell proliferation during embryonic development and bonehomeostasis in the adult. BMPs are promising molecules for tissue engineering and bone therapy. Thepresent review discusses this family of proteins, their structure and biological function, their therapeuticapplications and drawbacks, their effects on mesenchymal stem cells differentiation, and the cellsignaling pathways involved in this process.

� 2014 Published by Elsevier Inc.

mal stemor; FGFs,sforming

Introduction

The mechanism of bone tissue mineralization is complex event,involving pre-osteoblastic cells, physico-chemical events and aframework constituted by molecules present in the organic matrix[1,2]. Matrix proteins also play an important role in this process, asregulatory and/or nucleating factors in the deposition of hydroxy-apatite crystals, formed by calcium and phosphate ions present inplasma and extracellular fluids, in the space between the collagenmolecules [3]. Reparative regeneration occurs when tissues are lostdue to injuries or diseases, therefore, bone defect repair constitutesan adequate model to study bone regeneration. Unlike fractures,bone defects are less prone to mechanical factors and vascularsystem obstruction. Johner et al. [4] evaluated the regenerationof bone defects of rabbit tibia, observing that bone formationstarted within a few days without prior osteoclastic bone resorp-tion, with this regeneration being dependent not only on the size

of the defect, but, also, on the activity of pre-osteoblastic cells,and mechanisms which regulate their proliferation, differentiationand function.

In fact, it is widely known that different growth factors actlocally to modulate bone formation by stimulating pre-osteoblastsproliferation and activity [5]. A number of bone-derived growthfactors have been isolated and characterized from bone matrix,such as Bone Morphogenetic Proteins (BMPs),1 which display mito-genic, differentiating, chemotactic, and osteolytic activities, allowingthese molecules to act as potential determinants of local boneformation. BMPs are potent mediators of cell proliferation andmesenchymal stem cells (MSCs) differentiation, which have beenshown to be essential molecules involved in bone repair. In this

wth andrulation;mitogen

PRs, BMPcrophage

induced

ceFiGDBarecop

A.C. Carreira et al. / Archives of Biochemistry and Biophysics 561 (2014) 64–73 65

review, we discuss BMPs structural and functional aspects, theirsignaling pathways, their involvement in the differentiation of MSCsinto osteoblasts and their possible rople as therapeutic agents.

Growth factors related to bone formation and repair

Osteogenesis involves migration and mitosis of mesenchymalstem cells (MSCs), as well as their differentiation into osteoprogen-itor cells, and their differentiation and maturation into osteocytes(Fig. 1). The growth factors involved in the biological events ofbone and other connective tissues formation and repair may begrouped, according to their biological activities, into: Insulin-likeGrowth Factors (IGFs), platelet-derived growth factor (PDGF),Fibroblast Growth Factors (FGFs), Epidermal Growth Factor (EGF)and proteins from the Transforming Growth Factor Beta (TGF-b)superfamily (Table 1) [6,7].

The TGF-b family includes activins, nodal proteins, MullerianInhibiting Substance (MIS) and other Growth and DifferentiationFactors (GDFS), while BMPs represent the largest subgroup of thisfamily [8]. After identifying the role of BMPs in the initiation ofendochondral ossification [9,10], its role in embryonic develop-ment and cellular function has been extensively studied, withabout 20 BMP family members being characterized, highlightingthe role of this group as one of the main growth factors relatedto bone repair [11].

Bone Morphogenetic Proteins (BMPs)

In 1965, Urist demonstrated that demineralized, lyophilizedsegments of bone were capable of inducing new bone formationwhen implanted into ectopic sites, namely, rabbit muscle pouches[12]. In 1971, proposing the name ‘‘Bone Morphogenetic Protein’’[13]. In 1972, Reddi and Huggins showed that demineralized bonematrix is also capable of inducing bone formation in ectopic sites[14]. Since then, several BMPs were isolated, with these low molec-ular weight bone glycoproteins being proven to be the responsible

Fig. 1. Osteoblast differentiation. Temporal pattern of growth factors and trans

for promotion of this ectopic bone formation [15]. BMPs are syn-thesized by osteoprogenitor cells, osteoblasts, chondrocytes andplatelets [16,17] but their production is not restricted to bone,since they also play an essential role in development cell functions.

These proteins play a critical role in the development of manycell types in various tissues, acting in cell proliferation anddifferentiation, tooth morphogenesis, organogenesis, embryonicdevelopment, apoptosis, chemotaxis and repair of a wide varietyof tissues [18,19], in addition to glucose homeostasis and modula-tion of iron homeostasis (Table 2). BMPs induce endochondral/intramembranous ossification and chondrogenesis, by inducingmesenchymal stem cells differentiation towards the osteoblasticlineage [20], being critical for maintenance of skeletal integrityand in bone fracture healing.

BMP classification and structure

To date, around 20 different human BMPs have been found andgrouped into subfamilies, based on their sequence similarity andknown functions, even though not all members are trulyosteogenic (Table 2). Thus, BMP1 does not belong to the TGFbsuperfamily, being a metalloprotease that cleaves the C-terminusof procollagen I, II and III and being capable of inducing cartilageformation in vivo [21].

TGF-b superfamily proteins are classified according to theirprotein sequence similarity in humans (Fig. 2) and other species.The BMP family may be divided into four subfamilies accordingto their amino acid sequence (Table 2 and Fig. 2): (a) BMP2 and4 (80% homology); (b) BMP3, BMP3B (GDF10); (c) BMP5, 6, 7, 8aand 8b (78% homology); (d) GDF5, 6, 7 [22]. Curiously, BMP3 andBMP13 act either as a BMP negative regulator or as an inhibitorof bone formation, respectively [23].

BMPs are dimeric molecules, constituted by about 120 aminoacids, including seven conserved cysteine residues, from whichsix are highly conserved, comprising a cysteine knot motif linkedby three intramolecular disulfide bonds (Fig. 3). Another cysteineis involved in stabilization of the dimer through an intermolecular

cription factors expression during post-fracture osteoblast differentiation.

Table 1Extra or intracellular messengers involved in bone development and repair (Adapted from [38]).

Function Factor Process- in vivo and in vitro effects

Extracellular messengerOsteogenic factor BMP2 Initiates the bone formation and healing and induces the expression of other BMPs

BMP4 In vivo and in vitro osteochondrogenic factorBMP7 In vivo and in vitro osteogenic factor; active in mature osteoblasts

Mitogenic and osteogenic factor FGFb Gene mutations produce chondrodysplasia and craniosynostosis; stimulates expression of Sox9PDGFs Stimulates proliferation of osteoblasts, chondrocytes and mesenchymal cellsIGF-I, II Stimulates the growth plate, endochondral ossification and bone formation and osteoblastWnt Crucial to the proliferation of osteoprogenitors; can inhibit osteoblast maturationTGF-b Induces osteoblastic differentiation in undifferentiated cells; inhibits osteogenesis in committed cells

Osteochondrogenic factor Ihh Essential role in plaque formation and endochondral growth; can induce the PTHrP expressionPTHrP Induces chondrogenesis and suppresses hypertrophy

Epithelial factor EGF Stimulates proliferation of epithelial cells

Intracellular messengerSignal transduction in osteogenesis MPKs Essential for the regulation of intracellular signaling of osteogenic factors

PKA/CREB Transduction of osteogenic signalEarly osteogenic transcription factor Runx2 Main regulator during initial phase of osteogenesisOsteogenic transcription factor Dlx5 Induces maturation of osteoblasts and inhibits the osteocytes formation

b-catenin Essential role in the Wnt signal transduction and negatively regulated by GSK3bMsx2 Induces proliferation of precursor cells; the response depends on Dlx5

Late osteogenic transcription factor Osterix Main regulator during late osteogenesis; inhibits chondrogenesis

66 A.C. Carreira et al. / Archives of Biochemistry and Biophysics 561 (2014) 64–73

cystine bridge [24,25]. Some members of the BMP family, namely,GDF3, GDF9 and GDF9B, may form non-covalent dimers due to thelack of this cysteine [26].

BMPs are synthesized inside the cell as large and inactiveprecursors, carrying an N-terminal hydrophobic signal peptide(50–100 amino acids), which directs the protein to the secretorypathway, a pro-domain that mediates proper folding and aC-terminal mature peptide [27] (Fig. 3). These proteins also displaysites for N- and O-glycosylation, which increase their stability andhalf-life in the body and determine the specificity of receptorcoupling [28,29]. After dimerization, a prerequisite for boneinduction, the pro-domain is proteolytically cleaved at a consensusArg-X-X-Arg region to generate mature and active homodimers orheterodimers [30] (Fig. 4). BMP2 and BMP4 also possess a Furinconsensus motif at a secondary cleavage site [31,32]. While BMPsare active as a homodimer or heterodimer, some studiesdemonstrate increased potency of several BMP heterodimers,relative to BMP homodimers, both in vitro and in vivo, as shownfor BMP4/7 [33,34], BMP2/6 [35], BMP2/7 [36], BMP15/GDF9[36], and BMP2b/7 [37].

BMPs present widely recognized roles in bone formation duringmammalian development, exhibiting versatile functions. In bone,osteoprogenitor cells, osteoblasts, chondrocytes, platelets, andendothelial cells, produce BMPs. The regulatory effects of BMPsdepend upon the target cell type, differentiation stage, local BMPsconcentration, as well as interactions with other secreted proteins[38]. The efficacy of BMPs in accelerating bone regeneration andfracture healing has been demonstrated in animal models andpreclinical trials. BMPs act in conjunction with other growthfactors, in a complex cell signaling system.

BMPs signaling

BMPs are regulated by intracellular and, also, extracellular andintracellular molecules which bind BMPs or components of theBMPs pathways. Extracellular and intracellular regulators, whichact as antagonists of the BMP signaling pathway, may bind to thereceptors or sequester BMP ligands, blocking signal transduction,thereby decreasing bone formation. Over 15 BMPs antagonistsare known and classified in four groups: (a) Neublastoma (Dan)family (Dan, PRDC/GRem2, Gremlin, Cerberus/Cer1, Coco/Dand5,Caronte, USAG-1, Sclerostin/SOST, Dante/Dte); (b) Twisted

gastrulation (Tsg), (c) Chordin family (Chordin, Ventroptin/Chordin-like-1/Neuralin 1, Chordin-like-2, Kielin, Nel, Crossvein-less2/BMPER, Brorin, Brorin-like, Noggin) and Follistatina and (d)FLRG. [39]. Noggin expression in osteoblasts is limited and itsfunction blocks the effects of BMPs in undifferentiated anddifferentiated cells of the osteoblastic lineage [40] inhibitsmembranous ossification and prevents limb development andchondrogenesis. Xnr3, Lefty, BMP3 and BMP15, a subgroup of theTGFb superfamily, have direct interaction with and also act asBMPs antagonists [41,42]. But, thus far, no molecule has beendescribed as a BMP agonist.

BMPs bind to a hetero-tetrameric complex transmembranereceptor composed by type I (BMPR-I) and type II (BMPR-II) serinethreonine kinase receptors (Fig. 5), which contain an N-terminalextracellular ligand binding domain, a transmembrane domainand an intracellular region. The type II receptors are constitutivelyactive and phosphorylated. The activated BMPR-II promotesrecruitment and phosphorylation of the type IA BMP receptor(BMPR-IA or ALK3) and of the type IB (BMPR-IB or ALK6) receptorupon oligomerization (BMP-induced signaling complex, BISC). Thisleads to activation of the Smad-independent pathway, involvingthe mitogen activated protein kinase (MAPK) [43,44]. However,both receptors are indispensable for both canonical Smad-dependent (TGF-b/BMP ligands, receptors and Smads) and non-canonical Smad-independent (p38 mitogen-activated proteinkinase pathway, MAPK) pathways. Upon phosphorylation, thetype I receptor recruits R-Smads, leading to activation of theSmad-dependent pathway (through Smads 1, 2, 3, 5 or 8) andregulates the transcription of several target genes [45,46].

In addition to Smad activation, BMP signaling has also beenreported via Ras/ERK/MAPK, in response to tyrosine kinase recep-tor pathway activation [47]. Modulation of the SH3 domain activityof R-Smad (AMSH) may regulates the MAPK pathway, sinceR-Smad binds Smad 6 and inhibits the interaction with BMPRI[9], or inhibits Smad activity through the decrease of Smad1 inthe nucleus by its phosphorylation [47]. Protein–Protein interac-tion of BRAM and XIAP and TAK1 (TGF-b activated kinase 1) couldactivate the MAPK pathway. XIAP is recruited by BMPRI to linkTAK1 and activates p38 and JNK kinases and NF-jB. BMPs can alsoactivate ERK, PI3 kinase, PK A (Protein Kinase A), PKC and PKDpathways. Different cytokines, such as EGF, IGF and FGF, allowthe interaction of MAPK signaling with other pathways. SeveralBMP gene family members have been proven to be targets of the

Table 2Characteristics of different BMP family members. Adapted from Carreira et al. [38].

BMP Names Functions Subcellularlocation

Gene expression Humanchromosome

Protein Expression Year ofdescription/PMID

BMP2 BMP2A,BDA2

Induces bone and cartilage formation Secreted Intestine; kidney; larynx; connective tissue; spleen;prostate; amniotic fluid; eye; heart; ganglia; cervix;placenta; lung; pancreas; mixed; brain; stomach;embryonic tissue; uncharacterized tissue; liver;bone; uterus; testis; mammary gland

20p12 Abundant in lung, spleen and colon andin low but significant levels in heart,brain, placenta, liver, skeletal muscle,kidney, pancreas, prostate, ovary andsmall intestine

1988(3201241)

BMP3A Osteogenin,BMP3

Negatively regulates bone density; antagonizes theability of certain osteogenic BMPs to induceosteoprogenitor differentiation and ossification

Secreted Bone marrow; eye; uncharacterized tissue;intestine; lung; pharynx; muscle; mixed; embryonictissue; prostate; blood

4q21 Adult and fetal cartilage 1988(3201241)

BMP3B GDF10 Regulates cell growth and differentiation in bothembryonic and adult tissues

Secreted Testis; brain; ear; lung; mixed; eye; embryonictissue; prostate; connective tissue; pineal gland;larynx; heart

10q11.22 Femur, brain, lung, skeletal muscle,pancreas and testis

1995(8679252)

BMP4 BMP2B,BMP2B,MCOPS6,OFC11,ZYME

Induces cartilage and bone formation; acts inmesoderm induction, tooth development, limbformation and fracture repair; acts in concert withPTHLH/PTHRP to stimulate ductal outgrowth duringembryonic mammary development and to inhibithair follicle induction. (by similarity)

Secreted:extracellularmatrix

Intestine; embryonic tissue; placenta; liver;stomach; brain; bone; mixed; prostate; vascular;eye; uncharacterized tissue; testis; mouth;mammary gland; lung; pancreas; ovary; heart; skin;connective tissue; lymph node; lymph; adiposetissue; salivary gland; kidney

14q22–q23 Lung; lower levels in the kidney.Present also in normal and neoplasticprostate tissues, and prostate cancercell lines

1988(3201241)

BMP5 – Induces cartilage and bone formation Secreted Placenta; heart; thymus; uncharacterized tissue;uterus; testis; connective tissue; ovary; mammarygland; embryonic tissue; lung; mixed; pancreas;liver; brain; muscle; trachea; prostate; eye;intestine

6p12.1 Lung and liver 1990(2263636)

BMP6 VGR, VGR-1 Induces cartilage and bone formation; proposed rolein early development

Secreted (bysimilarity)

Vascular; eye; embryonic tissue; intestine; blood;uncharacterized tissue; placenta; brain; ovary; lung;mixed; prostate; uterus; testis; umbilical cord;larynx; cervix; kidney; mammary gland; connectivetissue; mouth; parathyroid

6p24–p23 – 1990(2263636)

BMP7 OP-1 Induces cartilage and bone formation; may be theosteoinductive factor responsible for thephenomenon of epithelial osteogenesis; plays a rolein calcium regulation and bone homeostasis

Secreted Brain; uncharacterized tissue; placenta; eye; testis;mixed; kidney; prostate; muscle; heart; embryonictissue; intestine; mammary gland; stomach; ovary;mouth; thymus; trachea; connective tissue; uterus;skin; tonsil; lymph; lymph node; lung; ganglia;pancreas; bone

20q13 Expressed in the kidney and bladder;lower levels seen in the brain;expressed in the developing eye, brainand ear during embryogenesis

1990(2263636)

BMP8A OP-2 Induces cartilage and bone formation; may be theosteoinductive factor responsible for thephenomenon of epithelial osteogenesis; plays a rolein calcium regulation and bone homeostasis. (bysimilarity)

Secreted Skin; thyroid; mammary gland; brain; testis;uncharacterized tissue; mixed; intestine; thymus;larynx; placenta; connective tissue; bone; spleen;embryonic tissue; stomach; prostate

1p34.3 – 2002(1460021)

BMP8B BMP8 Induces cartilage and bone formation; may be theosteoinductive factor responsible for thephenomenon of epithelial osteogenesis; plays a rolein calcium regulation and bone homeostasis (bysimilarity)

Secreted Skin; testis; liver; brain; mixed; ascites;uncharacterized tissue; intestine; pancreas; muscle;uterus; ovary; bone; bone marrow; spinal cord; eye;embryonic tissue; ganglia; prostate

1p35–p32 – 1992(12477932)

BMP9 GDF2 Potent circulating inhibitor of angiogenesis; couldbe involved in bone formation; signals through thetype I activin receptor ACVRL1 but not other Alks

Secreted Liver 10q11.22 – 2000(10849432)

BMP10 – Required for maintaining the proliferative activity ofembryonic cardiomyocytes by preventingpremature activation of the negative cell cycleregulator CDKN1C/p57KIP and maintaining therequired expression levels of cardiogenic factorssuch as MEF2C and NKX2-5; acts as a ligand forACVRL1/ALK1, BMPR1A/ALK3 and BMPR1B/ALK6,

Secreted (bysimilarity)

Ascites 2p13.3 – 1999(10072785)

(continued on next page)

A.C.Carreira

etal./A

rchivesof

Biochemistry

andBiophysics

561(2014)

64–73

67

Table 2 (continued)

BMP Names Functions Subcellularlocation

Gene expression Humanchromosome

Protein Expression Year ofdescription/PMID

leading to activation of SMAD1, SMAD5 and SMAD8transcription factors; inhibits endothelial cellmigration and growth

BMP11 GDF11 Secreted signal that acts globally to specifypositional identity along the anterior/posterior axisduring development; plays critical roles inpatterning both mesodermal and neural tissues andin establishing the skeletal pattern

Secreted(probable)

Eye; brain; uncharacterized tissue; mixed; skin;mammary gland; mouth; intestine; lung; uterus;amniotic fluid; muscle; blood; testis; liver;pancreas; tonsil; kidney; placenta; prostate;connective tissue; parathyroid; nerve; bone;embryonic tissue; thymus; umbilical cord; heart;adrenal gland; lymph node

12q13.2 – 1999(10075854)

BMP12 GDF7 May play an active role in the motor area of theprimate neocortex (by similarity)

Secreted (bysimilarity)

Kidney; brain; testis; uncharacterized tissue 2p24.1 – 1994(8145850)

BMP13 GDF6,CDPM2,MCOP4

Required for normal formation of bones and joints inthe limbs, skull, and axial skeleton. Plays a key rolein establishing boundaries between skeletalelements during development (by similarity)

Secreted(probable)

Embryonic tissue; mixed; testis; bone; brain;placenta

8q22.1 – 1994(8145850)

BMP14 GDF5,CDPM1,CDMP1,LAP4, OS5,SYNS2

Could be involved in bone and cartilage formation;chondrogenic signaling is mediated by the high-affinity receptor BMPR1B

Secreted Eye; embryonic tissue; lung; heart; bone; pituitarygland; brain; salivary gland; mixed; connectivetissue; prostate; skin; uterus

20q11.2 Expressed in long bones duringembryonic development

1994(8145850)

BMP15 GDF9B,ODG2, POF4

May be involved in follicular development; oocyte-specific growth/differentiation factor that stimulatesfolliculogenesis and granulosa cell (GC) growth

Secreted – Xp11.2 Ovary 1998(9849956)

BMP16 Nodal Essential for mesoderm formation and axialpatterning during embryonic development bysimilarity

Secreted (bysimilarity)

Embryonic tissue; testis 10q22.1 – 1999 (USPatent No.596503)

BMP17 Lefty1;LeftyB

Required for left-right axis determination as aregulator of LEFTY2 and NODAL

Secreted Embryonic tissue; joint; brain; mammary gland;lymph node; mixed; uncharacterized tissue; spleen;pancreas; testis; bladder; intestine; lung; liver; skin

1q42.1 Evidence at transcript level 2000 (USPatent No.6027917)

BMP18 Lefty2;LeftyA

Required for left-right (L-R) asymmetrydetermination of organ systems in mammals. Mayplay a role in endometrial bleeding

Secreted Embryonic tissue; testis; mixed; brain 1q42.1 Evidence at transcript level 2000 (USPatent No.6027917)

GDF1 DORV,DTGA3

May mediate cell differentiation events duringembryonic development

Secreted – 19p12 Brain 1990(1704486)

GDF3 KFS3,MCOP7,MCOPCB6

Negatively and positively control differentiation ofembryonic stem cells; role in mesoderm anddefinitive endoderm formation during the pre-gastrulation stages of development

Secreted(probable)

Uncharacterized tissue; testis; embryonic tissue;kidney; epididymis; mixed

12p13.1 – 1993(8429021)

GDF8 MSTN,myostatin

Acts specifically as a negative regulator of skeletalmuscle growth

Secreted Heart; brain; connective tissue; lung; eye;embryonic tissue; muscle; placenta

2q32.2 – 1997

GDF9 – Required for ovarian folliculogenesis; promotesprimordial follicle development. Stimulatesgranulosa cell proliferation; promotes cell transitionfrom G0/G1 to S and G2/M phases

Secreted Testis; brain; mixed; kidney; mammary gland; liver;adrenal gland; ovary

5q31.1 Oocytes of primary follicles 1993(8429021)

GDF15 MIC-1,MIC1, NAG-1, PDF,PLAB, PTGFB

May be involved in follicular development; oocyte-specific growth/differentiation factor that stimulatesfolliculogenesis and granulosa cell (GC) growth

Secreted(probable)

Placenta; skin; mixed; lung; intestine; kidney;prostate; liver; pancreas; muscle; uncharacterizedtissue; stomach; connective tissue; uterus; trachea;eye; brain; ascites; embryonic tissue; amniotic fluid;blood; testis; mammary gland; ovary; lymph node;lymph; vascular; umbilical cord; heart; cervix;spleen; bladder; bone; parathyroid

19p13.11 Highly expressed in placenta, withlower levels in prostate and colon andsome expression in kidney

1997(9139826)

68A

.C.Carreiraet

al./Archives

ofBiochem

istryand

Biophysics561

(2014)64–

73

Fig. 2. TGF-b superfamily proteins based on the alignment of mature domains using ClustalW/X.

Fig. 3. Schematic diagram of BMP2 topology (adapted from [80]). The monomer isstabilized by three disulfide bonds represented by lines between two sulfurmolecules (‘‘S’’). The cysteine knot constitutes the monomer core from which fourstrands of anti-parallel b-sheets (b1–b9) emanate, forming two ‘‘finger’’ projections,and the a-helix located opposite to the cystine bond is perpendicular to thestructure axis formed by the b-sheets, creating the ‘‘wrist’’ structure.

A.C. Carreira et al. / Archives of Biochemistry and Biophysics 561 (2014) 64–73 69

SHH (sonic hedgehog) signaling pathway. BMP-dependent down-regulation of Shh is achieved by interfering with the FGF andWnt signaling activities, which maintain shh expression. FGF sig-naling has a critical role in cartilage and bone differentiation[48], and FGF and BMP interact during the skeletal bone pathwaydifferentiation. Evidence is available indicating that BMP down-regulation of shh is mediated by the respective down-regulationof the Wnt/b-catenin signaling pathway, which is normallyinvolved in maintenance of shh expression [49]. BMPs have alsobeen implicated as potential interactors of the Ihh/PTHrP (Indianhedgedog/Parathyroid hormone related protein) feedback loop[50]. All of these signaling pathways are, in one way or another,involved in steps of mesenchymal stem cell precursors, bone andcartilage cell proliferation and differentiation [51].

Several other pathways and factors are involved in BMP signal-ing, which is modulated by stringent regulatory mechanisms andby cross-talk with several pathways [52], such as BMP andTGF-b-activin, expression of endogenous TGF-b, BMP-Smad andNotch and STAT/LIF.

BMP and mesenchymal stem cells (MSCs)

Several different BMPs (BMP2, BMP4, BMP6, BMP7, BMP9,BMP12 and BMP13) have been shown to induce MSCs differentia-tion, while BMP3 induces MSCs proliferation [53]. Studies haveshown that pluripotential bone marrow mesenchymal stem cells(MSCs), as well as adipocytes, myoblasts, fibroblasts and neuralcells respond to BMPs [3].

As described above, BMPs produce their effects through interac-tion with two BMP receptors (BMPRs), which are activated bybinding to extracellular BMP molecules (Fig. 5). Mechanistically,Smad1/5/8 complex is the most pertinent player during MSCsdifferentiation, during which adipogenic and osteogenic fates arerigorously programmed. In fact, it has been observed thatBMPR-IA and BMPR-IB recruitment is sufficient to induce MSCsdifferentiation [54].

For osteogenic and adipogenic lineage differentiation, twotranscription factors are essential, namely: Runx2 and PPAR-c[55]. Runx2 is involved in osteogenesis regulation and PPAR-c isthe master regulator of adipogenesis displaying anti-osteoblasticdifferentiation effects. Increased expression of runx2 and ppar-cis correlated with down-regulation of other transcriptional factors,such as those coded by osterix and c/erb, which are also involved inthe osteogenesis process [56]. PPARc functions affect BMP2-induced osteoblastogenesis but this mechanism remains unclear.On the other hand, TGFb1 and BMP2 inhibit adipocyte differentia-tion from bone marrow MSCs through suppression of the PPARctransactivation function [65].

BMP2 treatment attenuated mRNA expression during MSCs adi-pogenesis and osteoblastogenesis-related genes [55].

BMPs induce MSCs adipogenesis by recruiting both Smad1/5/8and MAPK signaling [57], both of which involve PPARc activationvia zinc finger transcription factor Schnurri-2 and C/EBPa [58,59].On the other hand, Smad6 negatively controls this signaling byreducing both PPARc signaling and BMP-associated adipogenesis[57]. Conversely, MAPK signaling disruption leads to inhibition of

Fig. 4. Structural organization of BMPs. Full-length BMP consists of a signal sequence, a pro-peptide, and the mature region. The mature protein is secreted as a homodimerlinked by disulfide bonds. The C-terminal mature protein is proteolytically cleaved from the pro-domain at an R-X-X-R sequence by proteases before dimerization.

Fig. 5. BMP signaling pathways: representation of Smad-dependent and -independent pathways and their mechanism of regulation. BMPs initiate signaling from the cellsurface by binding two different types of receptors (type I and type II) that can activate the Smad cytoplasmic proteins. The heterodimeric formation of type I R and type II Rmay occur before or after BMP binding, which induces signal transduction through Smads. BMPs can also signal through SMAD-independent pathways via mitogen-activatedprotein kinase (MAPK): Erk (extracellular signal-regulated kinase), JNK (Jun N-terminal kinase) or p38 (p38 MAPK). Three classes of Smad proteins include receptor-regulatedSmads (R-Smads), such as Smad1, 5 and 8; the common mediator Smad (Co-Smad), Smad4; and inhibitory Smads (I-Smads), Smad6 and 7. Smad6 binds type I BMPR,preventing Smad1,5,8 activation (phosphorylation) Phosphorylated Smad1, 5, 8 complexes associate with intracellular Smad4 for translocation into the nucleus, binds totranscription factors, activating target genes transcription. The transcription factors activated by BMPs can be any of the following: AP-1, bZIP, RUNX, Fox, bHLH,Homeodomain, Sp1, nuclear receptors, or IRF-7. The co-activators that are utilized are: CBP/p300, SMIF, MSG1, or ARC105. Smad6 and 7 can suppress signals by preventingassociation of R-Smad with Co-Smad or R-Smad phosphorylation. BMPs may act in an autocrine and paracrine manner. Furthermore, specific antagonists (Noggin andChordin) are also expressed (adapted from [80]).

70 A.C. Carreira et al. / Archives of Biochemistry and Biophysics 561 (2014) 64–73

A.C. Carreira et al. / Archives of Biochemistry and Biophysics 561 (2014) 64–73 71

adipogenesis, in response to BMPs, by decreasing PPARc expres-sion [57]. In addition, some studies have identified BMP signalingmainly at the earliest stages of MSCs adipogenesis, concluding thatit represents a start point in this process [60,61]. Curiously, BMP4signaling seems to be present in brown adipose tissue [62,63], withits expression in white adipocytes inducing a brown adipocytephenotype, including insulin sensitivity and increased energyconsumption [64].

Similarly to adipogenesis, BMP2, 4, 6, 7 and 9 receptors activa-tion during osteogenesis involves both Smad1/5/8 and MAPKdownstream signaling [66,67]. Among them, BMP2 and 7 have beenextensively studied as stimulators of MSCs osteogenesis in bothin vitro and in vivo models [68,69]. It has been shown that BMPsinhibitors (Noggin and Gremlin) are able to impair bone formation[70,71]. Furthermore, some authors have found that BMP2 is bothnecessary and sufficient for osteogenic commitment in vitro [72].Curiously, there are some differences according to the source ofMSCs: while murine-derived MSCs show a well-defined osteogenicresponse to BMP signaling, human MSCs show a variable response,maybe by modulating the expression of the Noggin BMP antagonist[73,74].

Enhancement of osteoblast differentiation is mediated by acti-vation of Runx2 and other transcription factors, such as Dlx5[75]. BMP2 may be regulated by other BMPs in osteoblasts, andthe BMP-2/4 promoter contains Runx2-binding sequences, imply-ing a positive feedback loop for regulation of the BMP signalingduring osteogenesis [76]. In the context of BMP signaling-inducedMSCs differentiation, two variables may define the fate of the pro-cess: (i) dosage, since lower BMP2 concentrations direct MSCstowards adipocyte formation, while higher concentrations pro-mote osteogenic differentiation [77] and (ii) receptor type, sincesignaling through BMPR-IA, in general, induces adipogenic effects,while signaling via BMPR-1B induces an osteogenic response [54].

Regarding MSCs chondrogenesis differentiation, it is known thatsonic hedgehog signaling initiates this process, which induces BMPsignaling and directs MSCs differentiation into the chondrogeniclineage. Peptide growth factors are very important for chondrocytesproliferation and differentiation, improving the quality of cartilagerepair. Studies have shown BMP4 as a promising candidate forpromotion of MSCs chondrogenesis [78]. Moreover, BMP-4 seemsto be also effective for chondro-lineage differentiation [79].

BMPs and tissue engineering

BMPs are the most effective growth factors in improving heal-ing of non-unions, fractures, spinal fusions and dental implants,as discussed in a recent review [80]. However, BMPs must be asso-ciated to a delivery system in order to exert and maintain their bio-logical activity at the surgical site in a controlled fashion, avoidingsystemic diffusion. The use of scaffolds loaded with osteo-progen-itor cells and/or growth factors has gained a great interest [81,82].While several studies have investigated other growth factors, suchas bFGF (basic fibroblast growth factor), GM-CSF (granulocyte-macrophage colony stimulating factor), IGFs, and PDGF (platelet-derived growth factor) for potential use in bone reparative therapy[6,52], only BMPs are currently available for clinical use.

BMPs act at very low doses (ng or lg) in the tissue; however,milligrams of BMPs are necessary in the surgical site in order toachieve osteoinductive properties, due to their short half-life. Thiscreates a high demand for a very efficient model system to producehigh quantities of high quality BMPs for clinical application.

Recombinant BMPs

Despite the existence of well-established methods for BMPspurification from demineralized bone matrix, this process is known

for being extremely laborious and inefficient, with low proteinrecovery. Preparations should start from a minimum of 100 kg offresh and washed cortical bone, which must be free of bone mar-row residues. Since BMPs molecular weights range from 15 to30 KDa, partial purification of BMPs from bone typically generates57 mg of a set of BMPs per kg of fresh bone [83]. It is important tonote that after purification, further steps are still required to iden-tify the set of isolated BMPs by protein purification. If the goal isthe isolation of a specific native BMP, expected returns are evenlower, in the order of lg/kg of bone tissue. Another factor to con-sider in BMPs purification from bone is the source of donor bone(most commonly, of bovine origin because of the large quantitiesrequired) and the potential risks of allogeneic source material,which may limit its clinical applications [84]. These issues,combined with the low recovery rate and the huge effort requiredfor the purification procedure, plus its low specificity for a givenBMP, have stimulated the replacement of BMPs purification frombone tissue by Molecular Biology/Genetic Engineering techniques(cDNA cloning and expression) in the production of these proteins.

Recent studies have sought to produce recombinant proteins insearch for new potential regulators of relevant tissue engineeringprocesses – such as osteoblasts differentiation, for example – andunderstanding of the mechanisms involved in bone formation,focusing future therapeutic strategies, especially new drugs designand genetic therapies. In this sense, recombinant BMP2 and BMP7expression and purification from bacterial and mammalian expres-sion systems, may be of great interest because of the osteoinduc-tive properties of these proteins [85]. Nevertheless, it has beensuspected that the osteoinductive properties of recombinant BMPsmay be relatively small, when compared to those of BMPs obtainedfrom bone by purification procedures. A detailed analysis of thesedifferences between the recombinant BMPs and of those purifiedfrom bovine origin, based on calcium ions content and on radio-graphs appearance of bone treated with both proteins, showed thatectopical bone tissue maturation in rat muscles may be 10 timeshigher when purified BMPs are applied [86].

Several hypotheses have been raised to explain the functionaldifferences between BMPs produced by these two different meth-ods. Reports pointing to differences in aminoacids sequences ofBMPs obtained by genetic recombination are available [87]. More-over, it has been suggested that different morphogenetic proteinsact coordinately in bone healing, underlining the requirement forseveral recombinant BMPs, used simultaneously, to achieve goodclinical results. Yet, it is known that not all expression systemsare capable of promoting oriented post-translational modifica-tions, such as glycosylation, thus affecting the proteins spatialstructure and activity [88].

As previously discussed, recombinant proteins may representimportant tools for both studies on osteogenic proteins expressionand function, such as for a more efficient BMPs production for ther-apeutic purposes. However, since the production of recombinantproteins requires an expression system, the choice of the expres-sion system becomes the determinant factor for efficiency andquality of the final BMP molecules produced. Currently, despite awide range of options for expression systems, they may classifiedinto:

(i) Bacterial: Escherichia coli is the most representative andsuitable bacteria for large scale production of proteins, due to theirhigh rates of growth and protein production, and low maintenancedemands. Although prokaryotic systems have been used forexpression of BMPs [89,90], post-translational folding stages,disulfide bonds formation and protein subunits association areoften compromised, resulting in reduced biological activity of theprotein produced.

(ii) Yeast based: Pichia pastoris methylotrophic yeast is oftenused as an expression vector for protein production, displaying

72 A.C. Carreira et al. / Archives of Biochemistry and Biophysics 561 (2014) 64–73

high growth rates, even in simple and low cost culture media [91].Expression vectors may also be developed to secrete the protein ofinterest, simplifying the purification process. Another advantage ofthis system lies on the fact that yeasts are capable of glycosylatingand producing disulfide bonds. However, it has been observed thatBMP7 produced using this system displays a much lower biologicalactivity than that expressed in mammalian cells [88]. One possibleexplanation for this problem may be the absence of specificchaperones required for its correct folding in yeast. In addition,issues concerning proper glycosylation may also interfere.

(iii) Baculovirus/insect cells system (BEVS): Transfection ofSpodoptera frugiperda cells by baculovirus carrying the gene ofinterest is commonly employed. It is a very well-studied system,which, despite the lower yield at higher costs and maintenance,has the advantage of including the main post-translationalmodifications and requirements for the correct protein foldingand aggregation. Although the human BMP2 produced using thissystem has yielded similar levels of osteo-induction whencompared to that of the same protein produced in a mammaliansystem [92], BEVS has seldom been used for BMPs expression.

(iv) Mammalian cells: BMPs produced using the mammalianexpression system display the most similar characteristics, whencompared to the original endogenous BMPs produced by tissues,since this system enables all stages of post-translational proteinprocessing, thus yielding correctly modified recombinant humanproteins [91]. In the case of BMP7, this includes the generation ofhomo- and heterodimers linked by disulfide bonds, originated byproteolysis of the native 431 aminoacids protein, displaying all ofthe expected glycosylation sites. The most commonly used cellsfor BMPs production in mammalian expression systems includeCHO, BSC-1 and COS7.54 cells [93].

Recently, recombinant human BMP7 was expressed in an alter-native platform using a bi-cistronic lentiviral vector to transformhuman 293T cells, thereby yielding dimers displaying high biolog-ical activity both in vitro and in vivo [85]. Therefore, selection of theappropriate expression system directly depends on the applica-tions foreseen for the protein to be expressed, be it structure-function relationship studies, mechanisms of action of BMPsat the molecular level, or large scale production of BMPs fortherapeutic purposes.

Clinical use

Recombinant human BMPs are used for orthopedic applicationsand oral maxillofacial surgery [94,95], becoming increasinglyimportant as an adjunct therapy for treatment of certain musculo-skeletal disorders.

Despite the fact that rhBMP2 and rhBMP7 have been approved,by the FDA, for anterior lumbar inter-body fusion and long bonedefects (non-union, open tibial fractures), respectively, receivinga humanitarian device exemption for revision postero-laterallumbar operations and recalcitrant long bone unions, availabilityof randomized controlled clinical trials on their diverse use inmany bone repair applications are still limited. BMP2 had benefi-cial effects in non-unions and in tibial fractures posing a high riskof healing complications. A systematic review summarized 18clinical trials and explored the osteo-regenerative potential usingBMPs [80].

New applications of BMPs have been found for treatment ofpost-traumatic osteonecrosis of the femoral head. The use ofBMP7 with other surgical alternatives may include core decom-pression, osteotomy, non-vascularized and vascularized bonegrafting, all of which may be enhanced [96].

In 2001 and 2002 the FDA approved the use of, respectively,rhBMP7 (Eptotermin-a) and rhBMP2, (Dibotermin-a), commerciallydenominated OP-1 Implant� (BMP7), and InductOS� and InFUSE�

(BMP2). The recombinant BMP2 and 7 approved for human use wereproduced in CHO cells, with their main applications being: to pro-mote fusion of lower spine vertebrae (InFUSE�, 2002, USA), acutetibia fractures (InductOS�2002, EU) and tibia non-union (2001, EUand USA). In 2004 (USA), rhBMP-2 produced in CHO cells was placedon an absorbable collagen sponge and its application was approvedfor acute open tibial shaft fracture. This is the same active ingredientpresent in the InFUSE� product. In 2009, the EU authorized the useof Opgenra (rhBMP-7 produced in CHO cells) for postero-laterallumbar spinal fusion in adult patients with spondylolisthesis.

Clinical application is not a routine yet in areas such asDentistry, for several reasons, including: the need for large dosesto be effective in humans, the decrease in response with increasingpatient age, quick release and high cost. Current BMPs applicationsexpose these proteins to the scene of inflammatory processes,which are inherent to early stages of injury repair, when intenseprotease activity probably inactivates a considerable part of theBMPs. Controlled release systems should be used to protect BMPsat this early stage, rendering them available only later on, duringthe restorative phase. Understanding the cellular and molecularbases of BMPs signaling pathways and the development of appro-priate carriers should stimulate a great revolution in Healthcare,particularly in Dentistry and Medicine. Clearly, well-designed,blind and randomized clinical trials are needed to achieve effectiveimplementation of BMPs in medical and dental clinics.

Conclusion

BMPs are potent regulators of bone and cartilage formation andrepair, cell proliferation in embryonic development and bonehomeostasis in the adult. They represent promising molecules inTissue Engineering, acting to accelerate and increase osteo-integra-tion. Understanding BMPs characteristics and signaling may lead tonew therapeutic applications of these proteins. However, whileresearch should continue to focus on improving the use of BMPsin the current clinical applications, the ability to engineer bone tis-sue and restore injured or diseased skeletal tissues represents aunique opportunity for BMPs in the future. Combination of BMPswith other peptide growth factors is also expected to become partof the available tools for treatment of bone pathologies, as well asassociation of BMPs with gene therapy and induced pluripotentstem cells (iPSCs). However, further studies are required to betterunderstand the applications and limitations of the clinical use ofBone Morphogenetic Proteins.

Acknowledgements

We thank all researches and technicians involved in our BMPprojects over the years. The financial support of the following Bra-zilian Federal and São Paulo State research agencies was essentialfor generation of this review and development of the work carriedout in our laboratory, namely: National Bank for Economical andSocial Development (BNDES), National Council for Scientific andTechnological Development (CNPq), Coordination for the Improve-ment of Higher Education Personnel (CAPES), São Paulo ResearchFoundation (FAPESP), Rio de Janeiro State Foundation SupportResearch (FAPERJ), Research and Projects Financier (FINEP),Ministry of Science, Technology and Innovation (MCTI) and Minis-try of Health of Brazil/Department of Science and Technology(MS-DECIT).

References

[1] D.W. Sommerfeldt, C.T. Rubin, Eur. Spine J. 10 (Suppl. 2) (2001) S86–S95.[2] N. Bivi, R. Pacheco-Costa, L.R. Brun, T.R. Murphy, N.R. Farlow, A.G. Robling, T.

Bellido, L.I. Plotkin, J. Orthop. Res. 31 (2013) 1075–1081.

A.C. Carreira et al. / Archives of Biochemistry and Biophysics 561 (2014) 64–73 73

[3] J.M. Granjeiro, R.C. Oliveira, J.C. Bustos-Valenzuela, M.C. Sogayar, R. Taga, Braz.J. Med. Biol. Res. 38 (2005) 1463–1473.

[4] R. Johner, Helvetica Chirurgica Acta 39 (1972) 409–411.[5] S. Mohan, D.J. Baylink, Clin. Orthop. Relat. Res. (1991) 30–48.[6] J.R. Lieberman, A. Daluiski, T.A. Einhorn, J. Bone Joint Surg. Am. 84-A (2002)

1032–1044.[7] M.W. Long, Blood Cells Mol. Dis. 27 (2001) 677–690.[8] P.G. Knight, C. Glister, Reproduction 132 (2006) 191–206.[9] E. Gazzerro, E. Canalis, Rev. Endocr. Metab. Disord. 7 (2006) 51–65.

[10] R.J. Wordinger, A.F. Clark, Exp. Biol. Med. 232 (2007) 979–992.[11] D. Buser, S.T. Chen, H.P. Weber, U.C. Belser, Int. J. Periodontics Restorative

Dent. 28 (2008) 441–451.[12] M.R. Urist, Science 150 (1965) 893–899.[13] M.R. Urist, B.S. Strates, J. Dent. Res. 50 (1971) 1392–1406.[14] A.H. Reddi, C. Huggins, Proc. Natl. Acad. Sci. U.S.A. 69 (1972) 1601–1605.[15] K. Bessho, T. Tagawa, M. Murata, Clin. Orthop. Relat. Res. (1991) 226–234.[16] M. Pecina, S. Vukicevic, Int. Orthop. 31 (2007) 719–720.[17] J.B. Sipe, J. Zhang, C. Waits, B. Skikne, R. Garimella, H.C. Anderson, Bone 35

(2004) 1316–1322.[18] P. Ducy, G. Karsenty, Kidney Int. 57 (2000) 2207–2214.[19] S.N. Lissenberg-Thunnissen, D.J. de Gorter, C.F. Sier, I.B. Schipper, Int. Orthop.

35 (2012) 1271–1280.[20] P. Leboy, G. Grasso-Knight, M. D’Angelo, S.W. Volk, J.V. Lian, H. Drissi, G.S.

Stein, S.L. Adams, J. Bone Joint Surg. Am. 83-A (Suppl. 1) (2001) S15–S22.[21] E. Kessler, K. Takahara, L. Biniaminov, M. Brusel, D.S. Greenspan, Science 271

(1996) 360–362.[22] M. Nakashima, A.H. Reddi, Nat. Biotechnol. 21 (2003) 1025–1032.[23] A. Daluiski, T. Engstrand, M.E. Bahamonde, L.W. Gamer, E. Agius, S.L.

Stevenson, K. Cox, V. Rosen, K.M. Lyons, Nat. Genet. 27 (2001) 84–88.[24] S.J. Butler, J. Dodd, Neuron 38 (2003) 389–401.[25] D.L. Griffith, P.C. Keck, T.K. Sampath, D.C. Rueger, W.D. Carlson, Proc. Natl.

Acad. Sci. U.S.A. 93 (1996) 878–883.[26] W.X. Liao, R.K. Moore, F. Otsuka, S. Shimasaki, J. Biol. Chem. 278 (2003) 3713–

3719.[27] K. Miyazono, U. Hellman, C. Wernstedt, C.H. Heldin, J. Biol. Chem. 263 (1988)

6407–6415.[28] K. Heinecke, A. Seher, W. Schmitz, T.D. Mueller, W. Sebald, J. Nickel, BMC Biol.

7 (2009) 59.[29] J.W. Lowery, J.M. Amich, A. Andonian, V. Rosen, Cell. Mol. Life Sci. 71 (2013)

3165–3172.[30] D.B. Constam, E.J. Robertson, J. Cell Biol. 144 (1999) 139–149.[31] Y. Cui, R. Hackenmiller, L. Berg, F. Jean, T. Nakayama, G. Thomas, J.L. Christian,

Genes Dev. 15 (2001) 2797–2802.[32] C. Degnin, F. Jean, G. Thomas, J.L. Christian, Mol. Biol. Cell 15 (2004) 5012–

5020.[33] A. Suzuki, E. Kaneko, J. Maeda, N. Ueno, Biochem. Biophys. Res. Commun. 232

(1997) 153–156.[34] K. Kusumoto, K. Bessho, K. Fujimura, J. Akioka, Y. Ogawa, T. Iizuka, Biochem.

Biophys. Res. Commun. 239 (1997) 575–579.[35] E. Valera, M.J. Isaacs, Y. Kawakami, J.C. Izpisua Belmonte, S. Choe, PLoS ONE 5

(2010) e11167.[36] D.I. Israel, J. Nove, K.M. Kerns, R.J. Kaufman, V. Rosen, K.A. Cox, J.M. Wozney,

Growth Factors 13 (1996) 291–300.[37] S.C. Little, M.C. Mullins, Nat. Cell Biol. 11 (2009) 637–643.[38] R. Dimitriou, E. Tsiridis, P.V. Giannoudis, Injury 36 (2005) 1392–1404.[39] B. Bragdon, O. Moseychuk, S. Saldanha, D. King, J. Julian, A. Nohe, Cell. Signal.

23 (2011) 609–620.[40] E. Gazzerro, V. Gangji, E. Canalis, J. Clin. Invest. 102 (1998) 2106–2114.[41] E. Di Pasquale, A.H. Brivanlou, J. Biol. Chem. 284 (2009) 26127–26136.[42] L.W. Gamer, J. Nove, M. Levin, V. Rosen, Dev. Biol. 285 (2005) 156–168.[43] J. Zhang, L. Li, Dev. Biol. 284 (2005) 1–11.[44] A. Herpin, C. Cunningham, FEBS J. 274 (2007) 2977–2985.[45] B.L. Rosenzweig, T. Imamura, T. Okadome, G.N. Cox, H. Yamashita, P. ten Dijke,

C.H. Heldin, K. Miyazono, Proc. Natl. Acad. Sci. U.S.A. 92 (1995) 7632–7636.[46] K. Miyazono, Y. Kamiya, M. Morikawa, J. Biochem. 147 (2010) 35–51.[47] J. Aubin, A. Davy, P. Soriano, Genes Dev. 18 (2004) 1482–1494.[48] X. Du, Y. Xie, C.J. Xian, L. Chen, J. Cell. Physiol. 227 (2012) 3731–3743.[49] M.F. Bastida, R. Sheth, M.A. Ros, Development 136 (2009) 3779–3789.[50] E. Minina, H.M. Wenzel, C. Kreschel, S. Karp, W. Gaffield, A.P. McMahon, A.

Vortkamp, Development 128 (2001) 4523–4534.[51] A.H. Reddi, Tissue Eng. 6 (2000) 351–359.[52] P.C. Bessa, M. Casal, R.L. Reis, J. Tissue Eng. Regen. Med. 2 (2008) 1–13.[53] G. Chen, C. Deng, Y.P. Li, Int. J. Biol. Sci. 8 (2012) 272–288.[54] D. Chen, X. Ji, M.A. Harris, J.Q. Feng, G. Karsenty, A.J. Celeste, V. Rosen, G.R.

Mundy, S.E. Harris, J. Cell Biol. 142 (1998) 295–305.

[55] I. Takada, Y. Yogiashi, S. Kato, PPAR Res. 2012 (2012) 607141.[56] M.T. Valenti, U. Garbin, A. Pasini, M. Zanatta, C. Stranieri, S. Manfro, C. Zucal, L.

Dalle Carbonare, PLoS ONE 6 (2011) e20363.[57] K. Hata, R. Nishimura, F. Ikeda, K. Yamashita, T. Matsubara, T. Nokubi, T.

Yoneda, Mol. Biol. Cell 14 (2003) 545–555.[58] W. Jin, T. Takagi, S.N. Kanesashi, T. Kurahashi, T. Nomura, J. Harada, S. Ishii,

Dev. Cell 10 (2006) 461–471.[59] V. Krishnan, H.U. Bryant, O.A. Macdougald, J. Clin. Invest. 116 (2006) 1202–

1209.[60] R.R. Bowers, M.D. Lane, Cell Cycle 6 (2007) 385–389.[61] R.R. Bowers, J.W. Kim, T.C. Otto, M.D. Lane, Proc. Natl. Acad. Sci. U.S.A. 103

(2006) 13022–13027.[62] S. Herzig, C. Wolfrum, Biochim. Biophys. Acta 1831 (2013) 895.[63] R. Hao, L. Yuan, N. Zhang, C. Li, J. Yang, J. Pediatr. Endocrinol. Metab. 25 (2012)

233–237.[64] S.W. Qian, Y. Tang, X. Li, Y. Liu, Y.Y. Zhang, H.Y. Huang, R.D. Xue, H.Y. Yu, L. Guo,

H.D. Gao, Y. Liu, X. Sun, Y.M. Li, W.P. Jia, Q.Q. Tang, Proc. Natl. Acad. Sci. U.S.A.110 (2013) E798–E807.

[65] A.W. James, Scientifica 2013 (2013) 1–17 (684736).[66] Q. Kang, W.X. Song, Q. Luo, N. Tang, J. Luo, X. Luo, J. Chen, Y. Bi, B.C. He, J.K.

Park, W. Jiang, Y. Tang, J. Huang, Y. Su, G.H. Zhu, Y. He, H. Yin, Z. Hu, Y. Wang, L.Chen, G.W. Zuo, X. Pan, J. Shen, T. Vokes, R.R. Reid, R.C. Haydon, H.H. Luu, T.C.He, Stem Cells Dev. 18 (2009) 545–559.

[67] L.J. Dorman, M. Tucci, H. Benghuzzi, Biomed. Sci. Instrum. 48 (2012) 81–87.[68] J. Reid, H.M. Gilmour, S. Holt, J. R. Coll. Surg. Edinburgh 27 (1982) 228–232.[69] M. Varkey, C. Kucharski, T. Haque, W. Sebald, H. Uludag, Clin. Orthop. Relat.

Res. 443 (2006) 113–123.[70] S.W. Davis, S.A. Camper, Dev. Biol. 305 (2007) 145–160.[71] W. Zhu, J. Kim, C. Cheng, B.A. Rawlins, O. Boachie-Adjei, R.G. Crystal, C. Hidaka,

Bone 39 (2006) 61–71.[72] D. Noel, D. Gazit, C. Bouquet, F. Apparailly, C. Bony, P. Plence, V. Millet, G.

Turgeman, M. Perricaudet, J. Sany, C. Jorgensen, Stem Cells 22 (2004) 74–85.[73] D.L. Diefenderfer, A.M. Osyczka, G.C. Reilly, P.S. Leboy, Connect. Tissue Res. 44

(Suppl. 1) (2003) 305–311.[74] C. Chen, H. Uludag, Z. Wang, H. Jiang, J. Cell. Biochem. 113 (2012) 3672–3680.[75] M.H. Lee, Y.J. Kim, H.J. Kim, H.D. Park, A.R. Kang, H.M. Kyung, J.H. Sung, J.M.

Wozney, H.J. Kim, H.M. Ryoo, J. Biol. Chem. 278 (2003) 34387–34394.[76] N. Ghosh-Choudhury, M.A. Harris, J.Q. Feng, G.R. Mundy, S.E. Harris, Crit. Rev.

Eukaryot. Gene Expr. 4 (1994) 345–355.[77] E.A. Wang, D.I. Israel, S. Kelly, D.P. Luxenberg, Growth Factors 9 (1993) 57–71.[78] R. Kuroda, A. Usas, S. Kubo, K. Corsi, H. Peng, T. Rose, J. Cummins, F.H. Fu, J.

Huard, Arthritis Rheum. 54 (2006) 433–442.[79] N. Shintani, E.B. Hunziker, Arthritis Rheum. 56 (2007) 1869–1879.[80] A.C. Carreira, F.H. Lojudice, E. Halcsik, R.D. Navarro, M.C. Sogayar, J.M.

Granjeiro, J. Dent. Res. 93 (2014) 335–345.[81] P.V. Giannoudis, O. Faour, T. Goff, N. Kanakaris, R. Dimitriou, Injury 42 (2011)

591–598.[82] P. Janicki, G. Schmidmaier, Injury 42 (Suppl. 2) (2011) S77–S81.[83] T. Gao, T.S. Lindholm, A. Marttinen, M.R. Urist, Int. Orthop. 20 (1996) 321–325.[84] C.A. Kirker-Head, Adv. Drug Deliv. Rev. 43 (2000) 65–92.[85] J.C. Bustos-Valenzuela, E. Halcsik, E.J. Bassi, M.A. Demasi, J.M. Granjeiro, M.C.

Sogayar, Mol. Biotechnol. 46 (2010) 118–126.[86] K. Bessho, K. Kusumoto, K. Fujimura, Y. Konishi, Y. Ogawa, Y. Tani, T. Iizuka, Br.

J. Oral Maxillofac. Surg. 37 (1999) 2–5.[87] K. Bessho, N. Tanaka, J. Matsumoto, T. Tagawa, M. Murata, J. Dent. Res. 70

(1991) 171–175.[88] S. Kujala, T. Raatikainen, J. Ryhanen, O. Kaarela, P. Jalovaara, Arch. Orthop.

Trauma Surg. 124 (2004) 26–30.[89] P.C. Bessa, M.T. Cerqueira, T. Rada, M.E. Gomes, N.M. Neves, A. Nobre, R.L. Reis,

M. Casal, Protein Expr. Purif. 63 (2009) 89–94.[90] S. Long, L. Truong, K. Bennett, A. Phillips, F. Wong-Staal, H. Ma, Protein Expr.

Purif. 46 (2006) 374–378.[91] W.H. Brondyk, Methods Enzymol. 463 (2009) 131–147.[92] Y. Maruoka, S. Oida, T. Iimura, K. Takeda, I. Asahina, S. Enomoto, S. Sasaki,

Biochem. Mol. Biol. Int. 35 (1995) 957–963.[93] D.H. Lee, H. Suh, D.W. Han, B.J. Park, J.W. Lee, J.C. Park, Yonsei Med. J. 44 (2003)

593–601.[94] C. Papakostidis, G. Kontakis, M. Bhandari, P.V. Giannoudis, Spine 33 (2008)

E680–E692.[95] K.A. Hussein, I.E. Zakhary, D. Hailat, R. Elrefai, M. Sharawy, M.E. Elsalanty, J.

Oral Maxillofac. Surg. 71 (2013) 1107–1118.[96] M.A. Mont, L.C. Jones, T.A. Einhorn, D.S. Hungerford, A.H. Reddi, Clin. Orthop.

Relat. Res. (1998) S314–S335.